John
It can be bypassed when you just click on yes when it asks is this correct.
I've done it myself, quite sure I typed the correct thing, and only picked
up my error on checking before putting the file in the Dropbox.
Maybe there could be a second query if, when the indexer says yes, it asks
'are you sure?'
Dawn M
From: ryerson_index-bounce@xxxxxxxxxxxxx
[mailto:ryerson_index-bounce@xxxxxxxxxxxxx] On Behalf Of John Graham
Sent: Saturday, 7 January 2017 3:19 PM
To: ryerson_index@xxxxxxxxxxxxx
Subject: [ryerson_index] Death Dates more than a month behind publication
When I run the update each day, I see about 10 or 15 messages for entries
where the death date is more than 30 days prior to the publication date. I
generally accept them without investigation, because I know that the indexer
gets a warning message when the record is being entered, and has to
consciously accept the message - so surely the date must be correct.
However, two entries in a file today now have me wondering if there is a
fundamental problem in the indexing program which lets these errors through
un-noticed by the indexer.
This is the situation from today. There were three consecutive notices,
notice A with a death date of 22 Dec 2016; notice B with a death date of 1
Jan 2017; notice C with a death date of 2 Jan 2017.
The data as submitted had the death date for notice A correct; notice B was
submitted as 1 Dec 2016; notice C was submitted as 2 Dec 2016.
I'm guessing that the indexer was using the date picker to enter the date,
and because the date in notice A was genuinely Dec 2016, then the indexer
didn't notice it was the incorrect month and year when entering notice B,
and again for notice C.
But in any case, the indexer would have received a warning message that the
death date was more than a month before the publication date for both
notices B and C - and that message is there specifically to trigger alarm
bells so that the indexer has to think as to whether or not the date entered
is correct. It takes a conscious action to accept the message.
How on earth do these sorts of errors get through? Can anyone enlighten me?
Is there a method of entering the date which somehow bypasses the warning
message?
In the past, I have always accepted such dates (I also get a warning when
running the update) on the basis that an indexer couldn't possibly make two
mistakes in the entry (month and year), and then accept the warning message,
if all three were wrong. Seems I am the one who is wrong, to assume that.
This is problematic as I see probably 10 or 15 such messages in each day's
updates. I don't have time to hunt down the notice and check each one, I
just accept them because of the warning message that I know the indexer has
accepted. But how many of them are wrong?
Can anyone explain how such results could be obtained, short of the indexer
being on auto pilot (and as these three notices were in the first 5 for the
day, that's hardly likely.)
John
---
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus
